Predicted survival in patients with brain metastases from colorectal cancer: Is a current nomogram helpful?

To examine the clinical applicability of a new nomogram by comparing survival of patients with brain metastases from colorectal cancer treated with surgery and/or radiotherapy in the authors' institutions with nomogram-predicted median survival.

Retrospective analysis of 64 patients treated with comparable approaches and during the same time period as the patients in the nomogram study. Points were assigned for age, performance status, number and site of brain metastases, as required for nomogram use.

In 46 patients (72%), the observed survival was shorter than the predicted median. The median deviation was -1.4 months. The nomogram underestimated the survival of patients treated with radiosurgery/surgery by a median of 4.2 months, whereas it overestimated the survival of patients treated with whole-brain radiotherapy (WBRT) by a median of 2.1 months (p=0.0001). Nevertheless, all 5 patients with predicted median survival ≤3 months died within 3 months. Among 8 patients with predicted median survival >12 months, 6 (75%) survived for >12 months. Not all prognostic factors in the nomogram correlated with survival. In the multivariate Cox model, only performance status and number of brain metastases were significant, both with p=0.0001.

Despite differences in prognostic factors and survival of many individual patients, especially those with intermediate prognosis, the nomogram performed promising in poor- and good-prognosis patients. Evaluation of separate prediction tools for patients treated with WBRT and more aggressive local approaches appears warranted in order to minimize the influence of better local control of the brain metastases.